A100 40GB PCIe Price & Used Prices: Which Local LLMs Can It Run?
The A100 40GB PCIe has 40GB of HBM2 memory with 1,555 GB/s of bandwidth. The median asking price is about $4,899 on eBay. At Q4 quantization it runs models up to 32B parameters at roughly 54 tokens/s.
A100 40GB PCIe price history
eBay
Full A100 40GB PCIe specifications
What LLMs can the A100 40GB PCIe run?
Versions released since 2026 with up to 50B parameters| Model | 2-bit | 4-bit | 8-bit | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Fits | Speed t/s | Context | Fits | Speed t/s | Context | Fits | Speed t/s | Context | |
| GemmaGemma 4 E4B | ✓ | 256.8 | 128K | ✓ | 189 | 128K | ✓ | 112.5 | 128K |
| QwenQwen3.5 9B | ✓ | 216.1 | 128K | ✓ | 158.4 | 128K | ✓ | 93.8 | 128K |
| GemmaGemma 4 12B | ✓ | 157.4 | 32K | ✓ | 118.5 | 32K | ✓ | 72.3 | 32K |
| GemmaGemma 4 26B A4B | ✓ | 159.8 | 32K | ✓ | 143.8 | 32K | ✓ | 114.2 | 32K |
| QwenQwen3.6 27B | ✓ | 87.6 | 32K | ✓ | 61.5 | 32K | ✓ | 34.8 | 32K |
| QwenQwen3.8 27B | ✓ | 87.6 | 32K | ✓ | 61.5 | 32K | ✓ | 34.8 | 32K |
| GLMGLM 4.7 Flash | ✓ | 190.9 | 128K | ✓ | 173.6 | 128K | ✓ | 140.6 | 128K |
| GemmaGemma 4 31B | ✓ | 66 | 8K | ✓ | 48.5 | 8K | ✗ | — | — |
| QwenQwen3.6 35B A3B | ✓ | 187.5 | 128K | ✓ | 170.7 | 128K | ✓ | 138.7 | 8K |
Fits when weights + KV cache + runtime overhead is at or below usable memory, single card. "Offload" on MoE models means it runs with expert weights in system RAM (only attention layers and KV cache stay in VRAM); that speed assumes 70 GB/s RAM bandwidth. The context column is the largest tier that fits.
Prices by platform
Buy links may carry affiliate tags; they do not change the price you pay.
FAQ
How much VRAM does the A100 40GB PCIe have?
The A100 40GB PCIe has 40GB of HBM2 memory with 1,555 GB/s of bandwidth.
Can the A100 40GB PCIe run a 70B model?
Not on a single card. At Q4_K_M with an 8K context a 70B dense model needs about 43GB, more than the A100 40GB PCIe's 40GB.
How much does a A100 40GB PCIe cost?
The median asking price is about $4,899 on eBay. Figures are medians of live listings, refreshed every 6 hours.